Online Handwritten Signature Verification using DCT

ABSTRACT Online handwritten signature verification is among the most widely used forms of behavioral biometric method, wherein the user submits his signature, captured through a handheld device for authentication. Time series data, such as X and Y positions have acquired during the signing process. Movement of pen with respect to X and Y Coordinates are derived for the verification purpose. The aim of the research is to present the implementation and analysis of an online handwritten signature verification using Discrete Cosine Transformation (DCT) for feature extraction and a multilayer feed forward neural network has been used for handwritten signature verification. The signature is verified in DCT domain in order to enhance the difference between a genuine and its forged signature. The present work compares it to one of the implementations using DWT and shows that DCT based verification provide more accuracy than DWT. Keywords: Discrete Cosine Transform (DCT), Multilayer feed forward neural network, Online Handwritten Signature, Signature Verification.